Recovery course of photoreceptor damage in a patient with cat scratch disease neuroretinitis: 2.5-year observation using adaptive optics retinal imaging

Abstract

Purpose

To assess the course of retinal damage in neuroretinitis associated with cat-scratch disease from the acute treatment phase to long-term follow-up.

Observations

A 29-year-old female diagnosed with seropositive cat-scratch disease neuroretinitis. At the initial visit, her right eye (OD) presented with decreased visual acuity (best-corrected decimal visual acuity, 0.8) and a central scotoma, along with hyperemic swelling of the optic disc, serous retinal detachment, and stellate maculopathy. Treatment with oral clarithromycin and oral prednisolone, starting with 30 mg daily and then tapered, led to complete resolution of these findings within 9 months. An adaptive optics retinal camera (rtx1™) revealed a disrupted photoreceptor mosaic corresponding to a post-detached retinal area, patchy dark spots in the papillomacular bundle, and hyperreflective deposits corresponding to stellate maculopathy. The dark spots and hyperreflective deposits diminished within one year. While photoreceptor mosaic disruption, correlated with disruption of interdigitation zone on OCT, showed gradual improvement, abnormal findings persisted in the papillomacular bundle even at 2.5 years. At the final follow-up, OD cone density remained at 84.7% of OS cone density in the macular temporal region (5° from fovea) (OD: 14,625/mm², OS: 17,263/mm²) and 66.5% in the macular nasal region (5°) (OD: 12,237/mm², OS: 18,401/mm²). MP-3 microperimetry (10-2 grid), covering these areas, showed normal retinal sensitivity in both eyes at 2-year follow-up, with no focal defects in the OD.

Conclusion

Long-term persistence of photoreceptor damage, even after visual recovery, may occur in cat-scratch disease neuroretinitis even after systemic antibiotic and anti-inflammatory therapies. Adaptive optics imaging is a useful tool for careful observation.

1. Introduction

Cat-scratch disease (CSD) is a major zoonotic disease caused by Bartonella henselae. Humans are usually infected by cat scratches, bites, or licking, whereas cat fleas act as transmission vectors among cats. Typical manifestations of CSD are focal reactions at the infection site, followed by systemic reactions including fever, fatigue, headache, and lymphadenopathy. Ocular complications are atypical manifestations of CSD that present as Parinaud's oculogranular syndrome, neuroretinitis, or retinal vascular occlusion,¹ with a reported prevalence of 5 to 10%.² Among these, neuroretinitis is particularly well-known as CSD represents the leading cause of infectious neuroretinitis.³ Neuroretinitis is defined as the inflammation of the optic nerve and peripapillary retina, typically presenting as optic nerve swelling, serous retinal detachment, and macular star formation.³ CSD is generally accepted as a self-limiting disease, and most cases of CSD-associated neuroretinitis have a benign prognosis with⁴ or without⁵ treatment. Although the pathogenesis of CSD-associated neuroretinitis is not fully understood, the primary target of inflammation is assumed to be the optic disc.⁶ While, macular involvement, including serous retinal detachment and the macular star, is secondary to hyperpermeability of the peripapillary capillaries, which could lead to photoreceptor damage.

Adaptive optics (AO) retinal imaging enable in vivo observation of retinal structures at cellular level by compensating for ocular wavefront aberrations, of which lateral resolution reach to 2 μm.⁷ Using this technique, we can directly visualize the cone photoreceptor mosaic, which is expected to provide new insights into various retinal diseases.⁸ In this report, we present detailed findings of photoreceptor alterations using an AO retinal camera in a patient with CSD-associated neuroretinitis who was treated with oral antibiotics and steroids.

2. Case presentation

A 29-year-old woman was referred to the Department of Ophthalmology at Kobe City Eye Hospital with visual field abnormalities in her right eye. She had adopted a kitten from a local animal welfare group approximately two months before first visit and had been frequently scratched by the kitten after that. She had experienced intermittent fever and headache for one month later, which was accompanied by a visual field abnormality in her right eye, 8 days before first visit. The patient had no past medical history.

At the initial examination, her best-corrected visual acuity (BCVA) was 0.8 decimal (0.10 logMAR) in the right eye and 1.5 decimal (−0.18 logMAR) in the left eye. The intraocular pressure was 13 mmHg in the right eye and 12 mmHg in the left eye. Goldmann perimetry revealed a central scotoma and blind spot enlargement in the right eye, whereas the left eye showed no abnormalities. There were no inflammatory cells in the anterior chamber or posterior segment of either eye. Fundus ophthalmoscopy revealed optic disc swelling and incomplete macular star formation in the right eye. Spectral-domain optical coherence tomography (OCT) (Spectralis; Heidelberg Engineering, Heidelberg, Germany) revealed subretinal fluid extending from the optic disc to the macular area, and fluorescein angiography revealed optic disc leakage in the right eye. The clinical appearance of the right eye at the initial visit is presented in Fig. 1. Serology was positive for IgG (titer 1:256) and IgM (titer 1:80) antibodies against Bartonella henselae, but negative for toxoplasma gondii, syphilis, and human immunodeficiency virus. Her blood pressure was within the normal range. Based on these findings, the patient was diagnosed with CSD-associated neuroretinitis. She was treated with oral clarithromycin (initial dose, 400 mg daily) and oral prednisolone initiated with 30 mg daily and then tapered. Antibiotic and steroid therapies were continued for two months and six months, respectively. The subretinal fluid in her right eye resolved one week after treatment initiation, with BCVA improvement in the right eye to 1.5 decimal (−0.18 logMAR). The macular star disappeared four months after treatment, and complete recovery of the visual field was confirmed using Goldmann perimetry nine months after treatment.

Fig. 1.

Clinical Appearances in the right eye at the initial examinations. (A) Wide-field color fundus photograph shows optic disc swelling and incomplete macular star formation. (B) early to mid-phase fluorescein angiography shows optic disc leakage. (C) Goldmann perimetry shows a central scotoma and blind spot enlargement. (D) infrared reflectance and (E) optical coherence tomography images show subretinal fluid extending from the optic disc to the macular area. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

We used an AO retinal camera (rtx1™; Imagine Eyes, Orsay, France) to monitor the course of retinal damage after treatment initiation. The AO retinal camera captured a 4° × 4° square area with a single acquisition. In this case, nine AO images were obtained to cover a rectangular area of 20° horizontally and 4° vertically, centered on the macular area. Image alignment and montage were performed using i2k Retina software (DualAlign LLC, Clifton Park, NY, USA). One week after treatment initiation, AO images showed a disrupted photoreceptor mosaic corresponding to a post-detached retinal area, patchy dark spots in the papillomacular bundle (PMB), and hyperreflective deposits corresponding to stellate maculopathy. Both the dark spots and hyperreflective deposits resolved within one year. Although the photoreceptor mosaic disruption, which correlated with a disruption of the interdigitation zone on OCT, demonstrated gradual improvement, abnormal findings persisted within the PMB even at two and a half years post-treatment. The ellipsoid zone line on OCT almost fully recovered after 2.5 years (Fig. 2). Cone mosaic metrics were analyzed at 2° and 5° temporally and nasally from the fovea, respectively. This analysis was performed using software provided by the manufacturer (AOdetect mosaic, version 3.0; Imagine Eyes, Orsay, France) for AO images. The size of the regions of interest (ROIs) was 62 × 62 μm for a standardized axial length of 24.0 mm. Prior to density calculation, ocular magnification was corrected based on the patient's axial length using the built-in scaling function in AOdetect, which applies Bennett's formula.⁹ An ophthalmic imaging specialist reviewed the image quality and selected ROIs where the cone mosaic was clearly resolvable. Following initial automated cone detection, three researchers independently reviewed and adjusted the identified cones to correct automated segmentation errors, and the mean values of these measurements were used for analysis. Table 1 shows the cone mosaic metrics, including density, spacing, regularity, and percentage of hexagonal cells, during the follow-up period. The markedly low cone densities and unanalyzable results, particularly in the early phase, reflected impaired cone visualization using this imaging modality due to photoreceptor detachment, misalignment, or temporary loss of reflectivity, rather than a reduction in the density of the remaining cones. At the final follow-up (2.5 years post-treatment), the cone density in the right eye remained 84.7% of the left eye density in the macular temporal region (5° from the fovea) (OD, 14,625/mm²; OS, 17,263/mm²) and 66.5% in the macular nasal region (5°) (OD: 12,237/mm², OS: 18,401/mm²). Similarly, other quantitative metrics (spacing, regularity, and percentage of hexagonal cells) initially showed unstable results, including unanalyzable measurements in the early phase, but gradually stabilized over 32 months. Cone spacing decreased over time; however, at 32 months, it remained longer in the right eye than in the left eye, with the nasal 5° region showing longer spacing than the temporal 5° region. While cone regularity exceeded 90% at all measurement sites in the unaffected left eye at 32 months, it remained in the 80% range in the right eye except at temporal 5°, showing lower regularity at nasal 5° compared to temporal 5°. The percentage of hexagonal cells also stabilized over time, but remained lower in the right eye (30–40%) than in the unaffected left eye (40–60%) across all evaluated locations at 32 months. Retinal sensitivity measurements obtained using microperimetry (MP-3; Nidek, Gamagori, Japan) covering the cone density analysis areas showed normal retinal sensitivity in both eyes at the 2-year follow-up (Fig. 3).

Fig. 2.

Adaptive optics (AO) retinal imaging observations following treatment initiation. (A) The imaging area for obtaining AO images in this study is indicated by the white box on the color fundus photograph. The white dashed line indicates the scanning line for the optical coherence tomography (OCT) B-scan. (B) AO montage image with visual markers (yellow boxes) indicating the evaluated regions of interest (ROIs) for cone mosaic metrics analysis (T: Temporal, N: Nasal). (C) Paired OCT and AO images are presented at 1 week, 9 months, and 2.5 years, along with enlarged inset images of the papillomacular bundle region including the nasal 5° ROI. Images at 1 week show photoreceptor mosaic disruption corresponding to a post-detached retinal area. Yellow arrows indicate the boundary between healthy and disrupted areas in the temporal macula. Yellow arrowheads point to patchy dark spots in the papillomacular bundle, and a yellow open arrow indicates a stellate maculopathy-related hyperreflective deposit. While dark spots and deposits resolve at 9 months, residual photoreceptor mosaic disruption remain detectable at 2.5 years. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Table 1.

Cone mosaic measurements during the follow-up period.

	5° Macular Temporal	2° Macular Temporal	2° Macular Nasal	5° Macular Nasal
Cone density
1 week, Right Eye (/mm2)	8121	12364	N/A	3470
9 months, Right Eye (/mm2)	11494	20806	21649	10804
32 months, Right Eye (/mm2)	14625	26882	26423	12237
32 months, Left Eye (Fellow Eye) (/mm2)	17263	33036	28882	18401
32months, Right Eye to Left Eye (%)	84.7%	81.4%	91.5%	66.5%
Spacing
1 week, Right Eye (μm)	12.5	9.63	N/A	19.6
9 months, Right Eye (μm)	10.4	7.73	7.48	10.3
32 months, Right Eye (μm)	9.12	6.71	6.75	9.89
32 months, Left Eye (Fellow Eye) (μm)	8.33	6.06	6.50	8.12
32 months, Right Eye to Left Eye (%)	109.5	110.7	103.8	121.8
Regularity
1 week, Right Eye (%)	85.5	86.3	N/A	92.1
9 months, Right Eye (%)	86.8	86.1	91.3	87.0
32 months, Right Eye (%)	92.4	88.5	88.0	86.9
32 months, Left Eye (Fellow Eye) (%)	93.1	95.7	91.5	93.7
Percentage of hexagonal cells
1 week, Right Eye (%)	27.0	45.0	0	48.7
9 months, Right Eye (%)	36.3	43.3	43.7	31.3
32 months, Right Eye (%)	36.0	40.3	33.0	39.0
32 months, Left Eye (Fellow Eye) (%)	55.0	43.0	44.7	52

Fig. 3.

Retinal sensitivity measurements by microperimetry 24 months after treatment. Solid yellow circles mark temporal and nasal locations at 5° eccentricity from the fovea. Dashed yellow circles mark locations at 2° eccentricity. There is no focal retinal sensitivity defects in the right eye. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

3. Discussion

In this case, we used AO imaging to perform a longitudinal cellular-level assessment of the retinal damage and recovery in a patient with CSD-associated neuroretinitis treated with systemic antibiotics and corticosteroids. Our findings highlight that the clinical resolution of optic disc swelling, serous detachment, and stellate maculopathy does not necessarily equate to complete restoration of the photoreceptor microstructure, revealing persistent subclinical damage years after the acute treatment phase.

The acute AO findings were consistent with the known pathophysiology of neuroretinitis and previous AO-based observations in other retinal diseases. The hyperreflective deposits correspond to lipid exudation within the Henle's layer, a characteristic of stellate maculopathy caused by leakage from peripapillary capillary inflammation.¹⁰ Widespread disruption of the cone mosaic¹¹ and patchy dark spots¹² likely represent photoreceptor outer segment misalignment or dysfunction, possibly due to serous retinal detachment and inflammatory reactions extending from the optic nerve head. Previous studies of various retinal diseases have reported a correlation between photoreceptor mosaic disruption visualized on AO and outer retinal band defects on OCT.¹³ In our patient, the observed spatial correspondence between the persistent AO mosaic disruption and interdigitation zone disruption supports this established relationship and is also applicable to CSD-associated neuroretinitis.

An important observation during the long-term follow-up was the topographical disparity in photoreceptor recovery. Cone density in the temporal macula showed better recovery to baseline levels comparable to the fellow eye (84.7%), with complete resolution of the cone mosaic disruption. However, the nasal macula, specifically within the PMB, exhibited an apparent cone density reduction (approximately 33.5% loss relative to the fellow eye at 2.5 years) and remaining cone mosaic disruption. These comparisons rely on the metrics that represent cones visualized and detected by the flood-illuminated adaptive optics system (rtx1™) and its automated software. One of the limitations of this modality is that cones with detachment, misalignment, or lost reflectivity may not be counted, even if they are still alive, which is clear from the markedly low cone counts in the early phase. Therefore, these metrics should be used as a marker for photoreceptor structural integrity, rather than as an exact count of viable cones. In neuroretinitis, the optic nerve head is considered the primary inflammatory focus.⁶ This concept is supported by characteristic fluorescein angiography findings, which typically demonstrate prominent leakage from the optic nerve head without macular leakage,¹⁴ as also observed in our case. Based on this proposed pathophysiology, we hypothesized that the nasal macula may be more susceptible to damage than the temporal macula due to its closer anatomical proximity to the optic disc. Persistent damage to this specific region likely stems from multifactorial mechanisms involving the PMB, including prolonged serous detachment, exudative changes associated with macular star formation, and direct inflammatory diffusion from the optic disc.

Finally, a notable insight is the dissociation between structural photoreceptor damage and preserved visual function. Despite the persistent cone mosaic disruption and cone density reduction, the patient demonstrated normal retinal sensitivity in the corresponding areas on microperimetry and remained asymptomatic. This discrepancy is likely attributable to photoreceptor redundancy and functional reserves, suggesting that visual function can be maintained up to a certain threshold of cellular damage.¹⁵ This finding aligns with the conventional understanding that CSD-associated neuroretinitis generally has a favorable visual prognosis.^4,5 However, we should be aware that subclinical, irreversible photoreceptor damage may persist long-term, and AO imaging serves as a useful tool for carefully observing these occult changes.

CRediT authorship contribution statement

Reina Aoyama: Writing – original draft, Visualization, Investigation, Formal analysis, Data curation, Conceptualization. Daiki Sakai: Writing – original draft, Visualization, Investigation, Formal analysis, Data curation, Conceptualization. Shin-ichiro Ito: Writing – review & editing, Supervision, Conceptualization. Wataru Matsumiya: Writing – review & editing, Supervision. Midori Yamamoto: Writing – review & editing, Investigation, Formal analysis, Data curation. Kiyoko Gocho: Writing – review & editing, Supervision. Sunao Sugita: Writing – review & editing, Supervision, Conceptualization. Makoto Nakamura: Writing – review & editing, Supervision. Yasuo Kurimoto: Writing – review & editing, Supervision.

Patient consent

Written consent was obtained from the patient for publication of this case report.

Disclosures

None of the authors has any financial disclosures or conflicts of interest related to this topic.

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work, the authors used ChatGPT and Gemini in order to improve readability and language. After using these tools, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.

Funding

This study did not receive any funding or grants.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.